Gearspace.com - View Single Post - Kirchhoff-EQ: Ultimate EQ plugin with superb dynamic processing and vintage models
View Single Post
Old 4 weeks ago | Show parent
Here for the gear
menzi11's Avatar
Originally Posted by engmix ➡️
I also have Claro. I A/B'd them. Matched their parameters, HPF to full spectrum cuts and boosts. They sound different enough that it might be worth having both. I found Claro was more clinical (in a good way) but the Kirchhoff seemed a little punchier (in a good way as well). And then when you start messing with the internal bit depth and phase responses, it goes a step further into sonic differences. Claro has the added A.I. stuff that I have found useful certain tasks. But the added dynamic eq options on the Kirchhoff was a big bonus. And way more so than on FF. You have the added benefit of Ratio, Q, Attack, and Release. Where on FF it's basically threshold and that's it. In a way I think this new offering is more suited to be wondering if it puts FF out to pasture over Claro. On the flip side, I'm honestly suspect of EQ Hardware Modeling that "to me" it's questionable if the added Neve, SSL, and GML/Sontec etc...are just gimmicks. I would like to see the developer chime in on this. How did they come about doing the emulations. Or are they simply inspirations so to speak. Which is all fine and good. The emulations do give you a little bit of a different vibe. But to me it's not like having a Sontec at your fingertips. And this is the case on every emulation I've tried, not just these guys. I own a stereo set of true to schematic GML/Sontec hardware clones, they're not the same, period. But, it's fun to have, but will it feel more like a gimmick versus a serious tool??

I'm the lead developer of Kirchhoff-EQ, Thank you for all the comments and you are right I think I should write some notes about how circuit modeling works in Kirchhoff-EQ, I will also put it on our website so everyone can see it:

There two methods for circuit modeling. The key is how to treat the distortion of the circuit. If we consider nonlinear things in the circuit, some kind of numerical integration should be used(as many many simulations did). Numerical integration will bring some high-freq cramping(similar to the bilinear transformation used by some EQ), and oversampling is generally used to avoid this problem.

the another way is to treat all circuit devices as linear(as many many simulations did), which is acceptable because the distortion of most EQ hardwares are not much. this way can make us match the lineared protype more better. and do more tricks such as add a dynamic function on.

just to me personally, I like a preamp plugin using numerical integration but don't like a EQ plugin using numerical integration. the reasons are still "round-errors", "high-freq cramping" and "anti-aliasing". but for linear way, sometimes we need those distortions even if they're not much.

So both methods have their own advantages and disadvantages, I totally agree about "every emulation I've tried don't like real gear". My idea is at least in Kirchhoff-EQ, not to do a simulate of distortions, DC Noises.... (since users may already have many of "100%" accurate simulations). Instead, keep it as a EQ, use linear models and with our "Robust Nyquist-matched Transform" and "Psychoacoustic Adaptive Filter Topologies", give users "the tastes" and give max freedom to use all the filter types together to see what will happened. does the vintage models in Kirchhoff-EQ 100% like the real? I can't say that. but in plugin world, I have full confidence in all the vintage models in Kirchhoff-Eq.